During the depressed phase of major depressive disorder and bipolar disorder, glucose metabolism is abnormally elevated in the amygdala, a brain structure that is critically involved in associating experiential stimuli with emotional significance and in organizing the endocrine, autonomic and behavioral responses to emotional stimuli. The magnitude of this elevation correlated positively with blood concentrations of the stress hormone cortisol. This relationship may reflect two phenomena that have been established in studies of rodents: the amygdalas prominent role in mediating corticosterone (the rodent equivalent of cortisol) secretion during stressful conditions, and the direct enhancement of amygdala function by corticosterone. In response to viewing pictures of human faces showing sad expressions, the neural responses of the amygdala and the prefrontal cortex were abnormal in the depressed phase of both bipolar disorder and major depressive disorder. In both depressed and healthy control subjects the blood flow initially increased in the amygdala in response to sad faces. However, after repeatedly viewing the same sad faces, the amygdala stopped responding in the healthy subjects, but continued to respond in the depressed subjects. This abnormality was associated with abnormal responses of the prefrontal cortex to sad faces in the depressives, such that activity in the anterior cingulate and orbital cortex increased in healthy subjects, but decreased in depressed subjects. These latter regions play important roles in regulating emotional responses generally, including those mediated by the amygdala. During the past one year three functional imaging experiments have been conducted to pursue these findings, and a fourth has now been initiated as well. One involved an investigation of the neural basis for the catastrophic response to perceived failure which depressed patients show. This study uses fMRI imaging during a probabilistic reversal learning task to elicit and characterize these phenomena. This study demonstrated initially a significant behavioral difference between depressives and controls, and a manuscript is in preparation to describe these results. During the past year, the image acquisition for this protocol was completed, the results were analyzed, and a manuscript describing these data is in preparation. The second imaging experiment investigated the effects of antidepressant drugs on amygdala responses to emotionally valenced faces. The pilot study initially characterized the optimal stimulus parameters needed to elicit the amygdala responses. The study of the depressed subjects was initiated and now a total of 20 depressed subjects and 20 healthy controls have been imaged. Many of the depressed subjects have been reimaged following fluoxetine treatment. Another study completed this year shows that depressed patients and healthy controls differ in the processing of emotional stimuli in a manner that is automatic, below the level of conscious awareness. We imaged 24 depressed subjects, 22 controls, and 14 unmedicated, remiitted subjects with a history of MDD. When the faces were presented using a technique called backward masking so that the subjects are not explicitly aware of having seen the emotional face, in the individuals who had a history of MDD the amygdala responded to sad, but not happy faces. In contrast, the healthy controls'amygdala responded to happy, but not sad faces. When the depressed subjects were treated with antidepressant drugs and re-imaged, their pattern of amygdala response became more normal, so that they were able to respond to the happy faces, and ceased responded to the sad faces. These data show how automatic the emotional processing biases are in depression. They also show that these biases persist into remission in MDD cases who are not receiving antidepressant drugs. In contrast, it appears that antidepressant drugs normalize these processing biases. This paper has now been submitted for publication. During the past one year we initiated a study in individuals who were at high risk for depression based on having a close family member with depression. This will show whether the emotional processing biases revealed by this task may be an endophenotype that is present before development of illness. The third study examined the rates at which depressives and controls are able to habituate to amygdala responses during repeated stimulation. Nineteen healthy volunteers were studied during the task development for this protocol. The data are being analyzed. During the coming year depressed subjects will be entered to investigate abnormalities of neural regulation during emotional processing in depression. Two manuscripts are in preparation describing these results. The new and fourth functional neuroimaging study that was conducted involved examination of the effects of catecholamine depletion, induced using the catecholamine synthesis inhibitor, alpha-methyl-para-tyrosine (AMPT), on cerebral glucose metabolism and on the neural responses to emotionally valenced stimuli in both recovered cases with MDD and in healthy controls. The data acquisition was completed for this protocol during this past year. Fifteen clinically remitted patients with ahistory of MDD and 13 healthy controls were scanned under both the placebo and AMPT conditions. The behavioral results confirmed previous evidence that this manipulation increases depressive symptoms in the recovered MDD cases. In addition, they demonstrated and characterized for the first time a greater vulnerability of the recovered MDD cases to develop anhedonia (inability to experience pleasure) and anxiety during the depletion, and hypomanic symptoms the day following the study in the recovery phase. The two manuscripts in preparation for submission have been accepted for publication. This past year we initiated a study of individuals at high familial risk for MDD and are imaging them under catecholamine depletion. Several of these individuals have developed depressive symptoms under depletion, whereas healthy controls with no family history of depression do not develop depressive symptoms--only fatigue. This implies that there is a heritable vulnerability that is modulated by the catecholaminergic neurotransmitter systems, such that imbalances to this system give rise to depressive symptoms. The previously conducted PET study assessing the effects of pramipexole on cerebral glucose metabolism was completed and the results analyzed. A manuscript is in press in Biological Psychiatry to report that these data showed that the functional anatomical abnormalities present in unmedicated subjects with MDD also are evident in lithium treated bipolar subjects who are currently depressed. In addition, a manuscript reporting that pramipexole treatment resulted in prominent reductions of metabolic activity in these regions is nearly complete.